In many cellular communications systems, the access to radio resources is controlled by the radio network. When a wireless transmit/receive unit (WTRU) has data to transmit to the network, it acquires radio resource access before transmitting its data payload. To achieve this in a 3rd Generation Partnership Project (3GPP) network, for example, a WTRU must gain access to the random access channel (RACH). Access to the RACH is contentious and there are mechanisms to reduce the probability of collision, that is, when two WTRUs are accessing the resource simultaneously.
Procedures for random access include a preamble phase with power ramp-up followed by channel acquisition information and message transmission. Because of the contentious nature of the RACH, to avoid WTRUs holding the shared radio resource for a long time, and because there is no power control, relatively short message payloads are transmitted on the RACH, leading to a relatively small data rate. Therefore, the RACH is generally used for the transmission of short control messages. Typically, WTRUs demanding larger data rates would be configured by the network to use dedicated resources.
While the data rate provided by the RACH is sufficient for the transmission of short control messages typical of networks supporting mostly speech communications, it is inefficient for the transmission of data messages associated with non-real-time data services, such as internet browsing, e-mail, and the like. For these data services, the traffic is bursty by nature and long periods of inactivity may exist between successive transmissions. For some applications requiring frequent transmission of keep-alive messages, for example, this may result in an inefficient utilization of dedicated resources. Therefore, it may be advantageous for the network to use shared resources for data transmission instead. The difficulty however, resides in the low data rate offered by the existing RACH.
FIG. 1 shows RACH access with a shared enhanced dedicated channel (E-DCH) 100 in accordance with the prior art. A RACH access with E-DCH 100, hereafter “E-RACH”, may include a RACH preamble phase 102, initial resource assignment 104, collision detection and resolution 106, an E-RACH message part 108, and release of resources 110 or transition to other state.
A WTRU may operate in multiple states. The form of the E-DCH while a WTRU is in CELL_DCH state may not be optimal for use in CELL_FACH state. In the context of CELL_FACH, the E-DCH suffers from three main drawbacks. First, the hybrid automatic repeat request (HARQ) processes with acknowledge/negative acknowledge (ACK/NACK) signaling in a shared environment are inefficient, particularly when only one medium access control (MAC) protocol data unit (PDU) needs to be transmitted. A MAC PDU may include a MAC-e PDU, a MAC-i PDU, or any other type of MAC-level PDU that is passed on to the physical layer. Second, there are too many downlink (DL) control channels Furthermore, the enhanced-dedicated physical control channel (E-DPCCH) overhead is too high.
During the E-RACH message part, the retransmissions occur at a fixed interval in time. For example, the retransmissions are separated by 3 TTIs for a system using 10 ms transmission time intervals (TTIs) and 7 TTIs in a system using 2 ms TTIs. If not all HARQ processes are used then the E-RACH is under-loaded and inefficient. In addition, the power control loop must be maintained even in periods of non-transmission.
The HARQ retransmissions can be inefficient in the context of a shared radio link where not all of the HARQ processes are occupied. For example, a WTRU may gain access to a shared E-RACH to transmit a single MAC-e PDU. If the MAC-e PDU is small enough, it will be transmitted in a single HARQ process, such as HARQ process 1, for example. In the context of E-DCH, if a Node B replies with a NACK on the corresponding HARQ indicator channel (E-HICH), then a retransmission occurs in the next HARQ process 1. This may occur after 14 ms in a system using a 2 ms TTI and after 30 ms in a system using a 10 ms TTI. Thus every time a single MAC-e PDU is transmitted, ⅞ of the HARQ processes are unused in the case of the 2 ms TTI and ¾ of the HARQ processes are unused in the case of a 10 ms TTI. Unless the WTRU has a relatively large amount of data to transfer, it can be concluded that using the shared resources in such a way is wasteful. Therefore, it would be desirable to have a set of mechanisms for efficient use of the E-DCH on the uplink RACH.